Mapping of hydrophobic sites on the surface of myosin and its

Mapping of hydrophobic sites on the surface of myosin and its fragments. Julian Borejdo. Biochemistry , 1983, 22 (5), pp 1182–1187. DOI: 10.1021/bi0...
0 downloads 0 Views 792KB Size
1182

Biochemistry 1983, 22, 1182-1187

is somewhat anomalous and are presently exploring this point. In general terms these results agree with the fluorescence results of Klausner & Wolf (1 98 1). As discussed above, the application of simple solution theories to the determination of phase partition in this system is limited. Partition is not sufficiently defined by the effect of solute upon transition temperature. Furthermore, use of fluorescence as a measure of the transition is complicated by the aggregation state and relative fluorescence quantium efficiency of the probe in the two phases as well as by the algorithm used to determine transition temperature. Calorimetric studies typically report T,,, as defined here, while fluorescence studies define T,,, as the temperature of half-effect (Klausner & Wolf, 1980). Figure 5 suggests that such a definition would considerably alter the dependence of of T,,, upon N . Application of fluorescent amphiphiles as probes of domain organization requires a knowledge of both their absolute partition and the partition of fluorescence. Registry No. DMPC, 18194-24-6; DPPC, 63-89-8; DSPC, 81694-4; CloDiI, 84109-06-8; CIZDiI,84109-08-0; CI4DiI,84109-10-4; C16DiI,84109-1 1-5; C18DiI,41085-99-8; CmDiI,84109-13-7;C,,DiI, 84109-15-9.

References Ackman, R. G. (1981) Methods Enzymol. 72, 205-252. Bearer, E. L., & Friend, D. S. (1982) J . Cell Biol. 92, 604-6 1 5. Buhr, M. M., Carlson, J. C., & Thompson, J. E. (1979) Endocrinology (Baltimore) 105, 1330-1 335. Foster, M. C., & Yguerabide, J. (1979) J . Membr. Biol. 45, 125-146. Hildebrand, J. H., Pravsnitz, J. M., & Scott, R. L. (1970)

Regular and Related Solutions, Van Nostrand Reinhold, New York. Klausner, R. D., & Wolf, D. E. (1980) Biochemistry 19, 6199-6203. Klausner, R. D., Kleinfeld, A. M., Hoover, R. L., & Karnovsky, M. J. (1980) J . Biol. Chem. 255, 1286-1295. Mabrey, R. D., & Sturtevant, J. M. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 3862-3866. Melchior, D. L. (1982) Curr. Top. Membr. Trans. 17, 263-3 16. Melchior, D. L., & Steim, J. M. (1979) Prog. Surf. Membr. Sci. 13, 211-296. Melchior, D. L., Scavitto, F. J., Walsh, M. T., & Steim, J. M. (1977) Thermochim. Acta 18, 43-71. Pessin, J. E., & Glaser, M. (1980) J . Biol. Chem. 255, 9044-9050. Pringle, M. J., & Miller, K. W. (1979) Biochemistry 18, 3 3 14-3 320. Ranck, J. L., Mateu, L., Sadler, D. M., Tardieu, A., GulikKrzywicki, T., & Luzzati, V. (1974) J . Mol. Biol. 85, 249-277. Setlow, V. P., Roth, S., & Edidin, M. (1979) Exp. Cell Res. 121 (I), 55-56. Shimshick, E. J., & McConnell, H. M. (1973) Biochemistry 12, 2351-2360. Sims, P. J., Waggoner, A. S . , Wang, C.-H., & Hoffman, J. F. (1974) Biochemistry 13, 1707-1716. Sklar, L. A., Miljanich, G. P., & Dratz, E. A. (1979) Biochemistry 18, 1707-1716. Wolf, D. E., Kinsey, W., Lennarz, W., & Edidin, M. (1981) Dev. Biol. 81, 133-138. Wu, S. H.-W., & McConnell, H. M. (1975) Biochemistry 14, 847-854.

Mapping of Hydrophobic Sites on the Surface of Myosin and Its Fragments? Julian Borejdo*

ABSTRACT:

The interaction of the hydrophobic probe conjugated polyene fatty acid cis-parinaric acid (PA) with myosin and its fragments was studied by measuring the enhancement of PA fluorescence following binding. The measurements point to the presence of about 1.34 hydrophobic sites per mol of myosin or per mol of double-headed fragment heavy meromyosin and 0.65 site per mol of myosin subfragment 1 (S-1). The binding constants were all in the range of lo7 M-'. The S-1 isoenzyme containing alkali light chain 1 [S-1(Al)] bound strongly close to 1 mol of PA per mol of S-l(A1) while the isoenzyme containing alkali light chain 2 [S-l(A2)] bound no

PA. Isolated alkali light chains A1 and A2 bound PA very weakly. Mg-nucleotide binding to the S-1 active site had little effect on the stoichiometry or affinity of PA binding; binding of actin to myosin carrying the probe either in solution or in an organized system of myofibrils had no effect on the magnitude of probe fluorescence but dramatically increased the rotational relaxation time of bound PA. These data indicated that parinaric acid bound to the hydrophobic pocket formed between the 41-residue region at the N-terminal end of A1 and the heavy chain of S-1 and that the hydrophobic pocket is distant from both actin and the nucleotide binding sites.

x e question of the effective (surface) hydrophobicity of the contractile proteins has not received much attention in the past.

Although many hydrophobic residues are buried in the interior of native proteins, some hydrophobic groups may remain exposed at the molecular surface or in crevices; in fact, the nucleotide binding site of myosin may be such a crevice. It is conceivable that some hydrophobic groups are involved in the hydrolysis of ATP and in the process of force development in muscle. For example, hydrophobic groups may have their accessibility to the solvent changed as a result of orientational alteration of the cross-bridge attitude believed to occur during muscle contraction (Huxley, 1969; Huxley & Simmons, 1971). Further, binding of substrate or change in the area of the

From the Polymer Research Department, The Weizmann Institute of Science, Rehovot 76100, Israel, and the Cardiovascular Research Institute, University of California, San Francisco, California 94143. Receiued July 2, 1982. This work was supported by grants from the US.-Israel Binational Science Foundation and the Muscular Dystrophy Associations of America and by Grant 8 1-612 from the American Heart Association. *Correspondence should be addressed to this author at the Polymer Research Department, The Weizmann Institute of Science. He is an Established Investigator of the American Heart Association.

0006-2960/83/0422-1182$01.50/0

0 1983 American Chemical Society

BINDING O F A HYDROPHOBIC PROBE TO MYOSIN

actin-myosin interface may alter the accessibility to the hydrophobic groups at, or near, the nucleotide binding site or at the point of contact between the thin filament and the tip of the cross bridge. Therefore, the study of the surface hydrophobicity of contractile proteins can potentially furnish useful information about cross-bridge attitude and conformational changes, occupancy of the enzymatic site, and changes at the actin-myosin interface. A series of original investigations demonstrated the usefulness of the hydrophobic probes in the study of myosin conformation and nucleotide binding: thus, in an investigation using a macromolecular probe [8-anilino-1-naphthalenesulfonate (ANS)] which was believed to bind to hydrobhobic regions of proteins [e.g., see McClure & Edelman (1966)], Duke et al. (1966) found a small enhancement of the ANS fluorescence upon binding to native myosin and a very large enhancement upon binding to the enzyme modified by p(ch1oromercuri)benzoate (PCMB). Lim & Botts (1967) have examined the effects of temperature and aging on the fluorescence of the myosin-ANS complex, and Cheung & Morales (1969) have shown that the efficiency of transfer of the excitation energy in the UV region from myosin chromophores to bound ANS decreased with modification of the enzyme with PCMB. These ANS fluorescence changes were interpreted in terms of the presence of different conformational states of native and modified myosin. Later, also using energy transfer from myosin chromophores to bound ANS, Cheung (1969) for the first time found evidence that myosin suffers a reversible change in localized structure during ATP catalysis. The present work is concerned with additional aspects of the interaction between contractile proteins and hydrophobic probes. The probe used here, a conjugated polyene fatty acid, cis-parinaric acid (PA), has been shown by Sklar et al. (1975, 1977) to be a useful molecule for studying the reversible binding to proteins; advantages of cis-parinaric acid as a fluorescence membrane probe have been reported by Sklar et al. (1976). Most importantly, Kat0 & Nakai (1980) have recently demonstrated that the “effective” hydrophobicity of a variety of proteins as determined by fluorescence enhancement following binding cis-parinaric acid correlates very well with the classical parameters of hydrophobicity such as the hydrophobic partition coefficient, interfacial tension, and the emulsifying activity index. Such a correlation may not hold for ANS because it has been shown (Penzer, 1972) that the fluorescence enhancement following binding of this substance to biological structures does not necessarily identify the binding site as hydrophobic. It is shown here that myosin, at both high and low ionic strength, has 1.3 high-affinity binding sites for reversible binding of cis-parinaric acid. Heavy meromyosin (HMM) is also shown to contain a little more than one binding site while S-1 is found to bind on the average 0.65 mol of cis-parinaric acid/mol of S-1 (with the same affinity). S-1 isoenzymes containing AI and A2 alkali light chains [S-l(A1) and S-l(A2), respectivelyl were found to differ greatly in their ability to bind PA: S-l(A1) bound PA with 1:1 stoichiometry and high affinity while S-1(A2) bound PA very weakly, if at all. I Abbreviations: ANS, 8-anilino-1-naphthalenesulfonate; A l , alkali light chain 1; A2, alkali light chain 2; PCMB, p(ch1oromercuri)benzoate; bis(ANS), 5,S’-bis(8-anilinonaphthalene-1 -sulfonate);S-1, myosin subfragment 1 ; HMM, heavy meromyosin;BHT, butylated hydroxytoluene; PA, cis-parinaric acid; BSA, bovine serum albumin; LMM, light meromyosin; Nbsz light chain, 19 000 molecular weight subunit of myosin dissociated by treatment with 5,5’-dithiobis(2-nitrobenzoic acid); NaDodSO,, sodium dodecyl sulfate; DTT, dithiothreitol; DEAE, diethylaminoethyl; EDTA, ethylenediaminetetraacetic acid.

VOL. 22, NO.

5,

1983

1183

Isolated alkali light chains also bound PA very weakly. On the basis of these observations, it is suggested that the hydrophobic binding site is formed between the 41-residue polypeptide at the N-terminal end of the alkali light chain 1 and the heavy chain of S-1. The fluorescence energy transfer between S-1 tryptophans and the hydrophobic probe and the influence of both nucleotide and actin binding on the environment of the PA binding site and on the polarization of fluorescence of the bound fluorophore have also been examined. Materials and Methods Proteins. Myosin was prepared from the back muscles of rabbit by the method of Tonomura et al. (1966). HMM and S-1 were routinely prepared from myosin by chymotryptic digestion according to Weeds & Pope (1977). The two S-1 isoenzymes were isolated by the procedure of Weeds & Taylor (1975). Alkali light chains AI and A2 were prepared according to Holt & Lowey (1975), except that precipitation of the heavy chain was achieved by dialysis of the mixture of heavy chains and light chains against low ionic strength buffer containing 80 mM potassium phosphate (pH 6.0) and 0.1 mM DTT. The light chains were fractionated by chromatography on DEAE-cellulose columns (2.5 X 60 cm). They were applied to the above buffer and eluted with a linear gradient to 0.27 M potassium phosphate and 0.1 mM DTT. All proteins were assayed for purity by NaDodS04-polyacrylamide gel electrophoresis. A1 always had a small contamination of the Nbs, light chain. Myofibrils were obtained from glycerinated rabbit psoas muscle as previously described (Borejdo, 1982). Fatty Acid. cis-Parinaric acid was purchased from Molecular Probes, Junction City, OR. PA solutions were freshly prepared each time by dissolving in ethanol at a concentration of 3-4 mM; equimolar butylated hydroxytoluene (BHT) was added as an antioxidant. All solutions were purged with nitrogen; the excitation slit of the fluorometer was kept as small as possible (see below). These procedures ensured negligible decomposition of the fatty acid. The concentration of parinaric acid was determined by the absorbance at 3 18 nm (=74 000 M-’ cm-I). Spectroscopic Measurements. Absorption spectra were obtained on a Cary 118 spectrophotometer. Fluorescence measurements were made at 20 OC on a Hitachi Perkin-Elmer MPF-4 spectrofluorometer in the ratio mode. The excitation and emission monochromator slits were set at 2 and 8 nm, respectively. Polarization of fluorescence was measured in a homemade apparatus by using the mercury lamp excitation, the excitation light monochromator set at 325 nm, and the emission barrier filter with a 400-nm cutoff. The fluorescence enhancement following mixing of protein and cis-parinaric acid developed rapidly and then kept increasing very slowly with time. For consistency reasons, therefore, at each probe/protein molar ratio, the fluorescence intensity recording began immediately after mixing and continued for exactly 3 min afterward. The initial reading was taken as a measured value. The sample was slowly stirred throughout the experiment. Energy transfer efficiency was calculated from fluorescence quenching of the tryptophan residues by using the equation T = 1 - FIFOwhere F is the fluorescence of PA-protein conjugates at 330 nm when excited at 280 nm and Fo is the fluorescence of the protein alone. ATPase Measurements. ATPase assays were performed by a pH-stat method in a Radiometer titrigraph at pH 7.6 and at 25 OC. The Ca2+-stimulatedactivity assay solution (3 mL) contained 50 mM KCl, 5 mM CaCI,, and 2.5 mM ATP. The EDTA assay solution contained 0.6 M KCl, 1 mM EDTA,

1184

BOREJDO

BIOCHEMISTRY I

I

I

I

I

Table I: Affinity and Stoichiometry of PA Binding to Myosin and Its Fragments moiety

K (X10' M-')

myosin

1.29 * 0.37 1.01 f 0.42 1.65 f 0.88

HMM s-1

Number of experiments.

1

0

I

I

I 2 Myosin odded I

1

bM)

3

1

I

12

FIGURE1: Binding of the hydrophobic probe to myosin measured by the enhancement of fluorescence. (A) Titration of PA (2.5 fiM) with myosin. Here and in all other experiments, the excitation wavelength was 325 nm (slit 2 nm) and the emission wavelength 410 nm (slit 8 nm), unless otherwise indicated. No correction for tryptophan fluorescence was made. (B)Scatchard plot of the data shown in (A). v is the number of moles of PA bound per mole of myosin. Myosin was in high ionic strength buffer containing 0.6 M KC1-5 mM phosphate buffer, pH 7.0.

and 2.5 mM ATP. Actin-activated ATPase was measured at pH 7.2 in 50 mM KC1, 5 mM MgC12, and 2.5 mM ATP. F-Actin was added to a final concentration of 0.05-1.5 mg/mL. All assay solutions contained 0.3 mg of S-1. Results PA Binding to Myosin. The quantum yield of PA in an aqueous solution is less than 0.001 (Sklar et al., 1977), and it increases 13-fold when the probe binds to myosin. The shape of the absorption and emission spectrum of PA was otherwise unchanged after binding to myosin or to its fragments. The enhancement of PA fluorescence following complex formation was used to estimate the affinity and stoichiometry of PA binding to myosin and its fragments. Two methods of titration were used. In the first method (method I), the PA concentration was fixed (usually at 2.5 pM), and aliquots of myosin or fragments were added to obtain the binding curve. In the second method (method 11), the concentrations of myosin or fragments were held fixed, and aliquots of PA were added. Figure 1A shows fluorescence enhancement of a solution of PA when it is titrated with the myosin at high ionic strength (method I). The initial linear increase in fluorescence extrapolates to a value close to 0.68 mol of myosin/mol of PA, and there is little further fluorescence increase when the myosin/PA ratio exceeds 1.0. The Scatchard plot of the data of Figure 1A is shown in Figure 1B. It has been assumed that maximum fluorescence corresponds to 100% binding, i.e., that the concentration (c) of free PA is given by [PAIfr, = (1 Z/Zmax) [PAItotwhere I,,, is the maximum fluorescence intensity and [PA],, is the added (total) concentration of PA.

vmax

Na

1.34 * 0.12 1.10 f 0.03 0.65 f 0.09

14 5 15

Mean * standard deviation.

v in Figure 1B is the number of moles of PA bound per mole of myosin. The slope of the Scatchard plot is the binding constant. The straight-line plot of Figure 1B indicates a simple binding site. The horizontal axis intercept indicates that there are, in this case, on the average vmax = 1.45 binding sites per mol of myosin. There may be other binding sites of low affinity as suggested by curvature of the plot at high v values. In 14 experiments of this kind, the average value of the association constant for binding of PA to the high-affinity site in myosin was (1.29 f 0.37) X lo7 M-l. The mean number of binding sites was 1.34 f 0.12 (Table I). A similar picture was obtained when myosin was titrated with PA (method 11, not shown). Here, the free PA concentration was computed as [PAItot= ZVmx/Zmaxwhere I,,, was the maximum fluorescence intensity and V,,, was the maximum concentration of bound PA. In this method, V,,, had to be assumed known a priori. In most cases, all values where V,, > 1.5[M], (where [MI, was the added myosin concentration, usually fixed at 2.5 pM) could be rejected because then all, or most, of the v/c values were negative. In five titrations by method 11, the average value of the binding constant and the average number of binding sites were (1.16 f 0.48) X lo7 M-' and 1.2 f 0.05, respectively. In an attempt to see whether the PA binding site overlaps, at least partially, with the nucleotide binding site of myosin, PA was titrated with myosin at low and high ionic strength in the presence of MgADP and MgATP. Neither K, nor ,v was influenced by either MgATP or MgADP binding, indicating that the nucleotide and the PA binding sites were distinct. We next probed the question of whether the hydrophobic binding is affected by the state of aggregation of myosin. To this end, PA was titrated with myosin at low ionic strength. The mean values of K, and ,v determined from six such experiments were (1.60 f 0.28) X lo7 M-' and 1.35 f 0.1 1, respectively, not significantly different from the values obtained at high ionic strength. PA Binding to Myosin Fragments. The typical Scatchard plots for binding of PA to the double-headed HMM and to S-1 are shown in Figure 2A. The K, values for binding to either fragment were similar; the affinity of binding was as strong as in the case of myosin. The stoichiometry was on the average 0.65 f 0.09 for S-1 and 1.1 f 0.03 for HMM. These data are summarized in Table I. MgATP and MgADP had no effect on PA binding to S-1 or HMM. PA Binding to S-I(A1) and S - l ( A 2 ) . The fact that one molecule of PA bound to on the average 0.75 molecule of myosin and 1.5 molecules of S-1 suggested that the hydrophobic site was in fact located not on the heavy chain but on one of the light chains of the S-1 molecule. This is because A1 and A2 light chain isoenzymes are known to be present in rabbit skeletal muscle in roughly 3:2 ratio (Sarkar, 1972; Weeds et al., 1972). Binding of PA to both S-1 isoenzymes was examined to test this possibility. The typical Scatchard plots for binding of PA to S-l(A1) and S-l(A2) are shown in Figure 2B. The affinity of binding to S-l(A1) was essentially the same as to unfractionated S-1 while the amount bound was now closer to 1 mol of PA/mol of S-l(A1). In

B I N D I N G O F A H Y D R O P H O B I C P R O B E TO M Y O S I N C

6

1 .o

II B

A \

4 W

b 7 X

0

‘s 2

,

\To I

0.5

\\ 0

00

\

1

I

I

I

Y

FIGURE2: (A) Scatchard plots for binding of PA to S-1 (filled circles, solid line, bottom scale) and to HMM (open circles, dashed line, top scale). Fragments were in 80 mM KCI, 5 mM sodium phosphate buffer, pH 7.0, and 0.1 mM MgC12. PA concentration was fixed at 2.5 pM. (B) Scatchard plots for binding of PA to S-I(A1) (filled circles, solid line) and to S-l(A2) (open circles, dashed line). Fragments were in 80 mM KCI, 5 mM sodium phosphate buffer, pH 7.0, and 0.1 mM MgCI,. PA concentration was fixed at 2.5 pM. otherwise, conditions as were described in Figure 1.

contrast, PA bound very weakly to S-l(A2) (Figure 2B). PA Binding to Isolated Light Chains. The difference in PA binding to the S-1 isoenzymes could be due to binding either to the 41-residue difference piece at the N-terminal end of A1 or to the hydrophobic site involving the heavy chain which is present in S-I(A1) but not in S-l(A2). To distinguish between these possibilities, I have measured PA binding to the isolated light chains. To 2.5 p M PA in 80 mM KC1-5 mM phosphate buffer, pH 7.0, were added aliquots of 0.25 pM A I (which had a small contamination of Nbs, light chain) or A2. PA bound at least 100-fold weaker to either light chain than to unfractionated S-1. The significance of this finding is dealt with under Discussion. Energy Transfer from Tryptophan to PA. The spectral overlap between the absorption spectrum of PA and the emission spectrum of S-1 tryptophans should lead to considerable energy transfer between the chromophores, providing the distance between them is sufficiently small. We have chosen to follow the energy transfer by exciting the tryptophans at 240 nm, below the absorption limit of PA, and to follow the changes in the emission spectra of S-1 tryptophans. The change in the tryptophan fluorescence of S-1associated with the binding of PA is shown in Figure 3A. The weighted energy transfer efficiency for these data calculated according to the equation defined under Materials and Methods is shown in Figure 3B. The low transfer efficiency for S-1 (25% even at 3.5 mol of PA/mol of S-I, cf. Figure 3B) is in line with the proposal of Kat0 & Nakai (1980) that the transfer efficiency correlates well with the effective hydrophobicity of a protein. By their standards, S-1 is weakly hydrophobic, and so the efficiency of transfer was expected to be low. We have not attempted here to resolve the changes in transfer efficiency following the nucleotide binding because of the complicating effects of ATP on the emission spectrum of S-1 tryptophans. Such changes were previously detected by Cheung (1969) for the case of ANS and by Takashi et al. (1977) for bis(ANS) bound to myosin. Also, we have not attempted to decompose the observed total transfer efficiency into the contributions due to each donor separately because the measurement of the total transfer efficiency with just one type of acceptor does not yield enough information for the

Moles PA/Mole S-l

3: (A) Fluorescence emission spectra of S-1 in the presence of increasing concentrations of PA. The S-Iconcentration was fixed at 2.5 pM. The numbers refer to ratios of moles of PA per mole of S-1. The excitation was at 240 nm (slit 4 nm); the emission slit was 20 nm. (B) Energy transfer efficiency as a function of the molar ratio of PA to S-1. FIGURE

determination of the individual transfer efficiencies. PA Binding to Actin and Acto-S-1. Polarization of Fluorescence of Bound PA. Under conditions where actin existed in the F form, we saw very weak binding of PA. The mean K, was